Problem Set 1: Linear Regression Analysis

Research Scenario: A cognitive psychologist is interested in the Dunning-Kruger effect, the phenomenon in which people often rate their cognitive skills as being greater than they really are. He recruits a sample of 17 undergraduate students who are near the end of an entry-level macroeconomics course at his university. The students are asked to take a test worth 100 points, on which most material is based on upper-level economics concepts (in other words, it's hard!). Once they finish the test, the students are asked to rate their proficiency level in economics on a scale of 1-10, where higher ratings indicate higher self-perceived proficiency in economics. The question at stake is whether the students' proficiency level self-ratings predict their actual performance on a difficult economics test.

Proficiency Level Actual Score

Self-Rating (1-10) (0-100)

-----------------------------------------------------------------------

8 54

7 67

7 83

6 72

8 31

9 29

5 84

4 78

5 67

9 20

8 39

8 41

6 47

7 39

6 56

7 28

9 22

Q. SPSS output?

Q. Results ?

Q. Graph ?